00 / PORTFOLIO & ADVISORYReady for the next challenge

Making Sense of the Data Mess

Connecting hardcore code with artificial intelligence to build awesome things (that actually work in real life).

Explore My Services→Check My Journey
Marcel Souza
FIG. 01 — MARCEL SOUZADATA & ML ENGINEERING
01Harvard: Data Science Pro02MIT: AI Strategy03Python Whisperer04Neural Nets That Work05AWS Certified (Survivor)06Master of Big Data
01 / Who I AmData Scientist • Problem Solver
Marcel Souza Professional
FIG. 02 — PROFILEData Scientist • Problem Solver

The Brain Behind The Models

I don't just train models; I teach machines how to think better. With over 16 years on the engineering road, I treat Data Science like a giant puzzle: it's fun, challenging, and incredibly satisfying when the pieces just fit.

I'm the guy who translates mountains of numbers into ideas you can actually use. I've seen it all in tech—from servers crashing on a Friday to AI that looks like magic (but is really just good math and solid code). My superpower is turning complex problems into simple, scalable, and production-ready solutions.

Years Coding
16+
Live Models
50+
02 / 16 Years of Tech Adventures2008 — PRESENT

When Code meets Data

012008 — 2014

The Heavy Backend Era

It all started building massive systems in Java and C#. Pouring foundations that handle enterprise workloads, focusing on making sure the software doesn't break when people actually use it.

JavaSQLSystem Design
022015 — 2019

Diving into Big Data

The game shifted to data. Creating giant pipelines to process terabytes of info without choking. Figuring out how to tame rivers of data.

SparkHadoopPipelines
032020 — Present

Teaching Machines to Think

Mixing the madness of code with the elegance of statistics. Pushing AI models into the real world to solve problems that seemed impossible.

Python MagicNeural NetsLLMs
03 / What I DoADVISORY & EXECUTION

I don't sell hype. I deliver results.

SRV // 01→

Data Consulting

Full diagnosis of analytical maturity, data strategy design, and implementation of analytics and ML solutions aligned with business goals.

SRV // 02→

Production ML Models

From business problem to a model running in production. Forecasting, classification, NLP, computer vision — backed by architecture that actually works.

SRV // 03→

Talks & Workshops

Talks on data, AI, tech careers, and transitions. From corporate events to local communities. Technical depth with soul, without 200-slide decks.

SRV // 04→

Tech Career Mentorship

Personalized guidance to enter or grow in Data Science. Portfolio building, positioning, interviews, and insights from someone with 16 years on the road.

04 / SkillsCORE CAPABILITIES

What I master. For real.

Not a generic list of buzzwords. This is what I use to solve real problems.

SPEC // 01

Machine Learning

Supervised and unsupervised models, from data exploration to production deployment. I don't just train — I deliver.

Scikit-learnXGBoostTensorFlowMLflow
SPEC // 02

Data Analysis

Transforming raw data into actionable insights. Exploration, cleaning, visualization, and storytelling with data.

PandasSQLPower BIMatplotlib
SPEC // 03

Software Engineering

16 years of coding. Architecture, APIs, versioning, CI/CD. What makes the model leave the notebook and go into production.

PythonGitDockerREST APIs
SPEC // 04

MLOps & Production

Data pipelines, model monitoring, training automation. The engineering side of a data scientist.

AirflowFastAPIAWSCI/CD
SPEC // 05

Business Vision

Data without context is just numbers. I understand the business pain before writing the first line of code.

KPIsROIStakeholdersStrategy
SPEC // 06

Tech Communication

I write, speak, and teach tech with clarity and good humor. Technical content that real people understand.

LinkedInYouTubeTalksMentoring
05 / ContentESSAYS & NOTES

I write about what I live.

Articles, reflections, and analyses about data, tech careers, and the technology ecosystem. No fluff. With code when needed, with stories when it matters.

NO. 01/Data Science

Why your ML model won't go to production (and how to change it)

Most models die in the notebook. The problem isn't the algorithm — it's the lack of engineering.

Coming soon
NO. 02/Career

16 years as a dev and I decided to become a Data Scientist. Am I crazy?

The honest story of how (and why) a senior dev decided to start over — without romanticizing.

Coming soon
NO. 03/Opinion

The Northeast tech scene goes beyond the axis. And Petrolina is proving it.

On building a tech community where "there is no market" and why it's the biggest opportunity.

Coming soon